- opinion
- ai-safety
- blog-logistics
- updates
- guides
•
•
•
•
-
Are we cooked? | Reflections from 2026 MAIA-HSSIT Workshop.
.
-
Silent Semi-Killers | Extinction-level impacts of AI that are right under our nose.
Silent Semi-Killers are a group of underrepresented AI catastrophic outcomes that lead to major (but not directly extinction-level) effects on our society and are hiding in plane sight. They are often arise from a need to increase productivity in our society, and target systems that haven't/need to adjust to AI usage.
-
Opinion Piece | Alignment and Safety Benchmarks are Cope, and Bring Complacency
Anthropic, the king of putting "safety at the frontier", just dropped their newest and most powerful model to date. Yet with it comes questions about its true safety and alignment. The recent news motivated me to write some thoughts on "benchmarks" and the current state of safety from my POV. Many are using benchmarks to justify scaling, leading to complacency that harms the AI safety community as a whole.
-
Upcoming Blogs
Upcoming Posts/Ideas, here for accountability sake.
-
25-26 Winter Studies Recap, Causality and Causal Representation Learning
A primer on causality and causal representation learning(kinda)